MDDI 演讲稿 · 2026-01-14

刘佩芬政务次长在新加坡国立大学 AI 与数码转型高管硕士课程启动仪式上的开幕致辞

Opening Address by MOS Jasmin Lau at Launch of NUS Executive Master in AI and Digital Transformation on 14 Jan

Jasmin Lau · MDDI 政务次长 · 新加坡国立大学 AI 与数码转型高管硕士课程启动仪式

要点

  • AI 落地不是技术练习,而是领导力练习。Jasmin 列出 3 个领导力挑战:①深度理解组织、知道 AI 在哪里能创造真实价值(卫生部护士登记 100 步流程的例子——先「归零」再数字化);②评估能力、决定 build/borrow/buy 的混合配比;③理解 AI 对人的影响,知道「即便能用也不该用」的时刻。
  • 新加坡部长们也在去年底接受了两轮 AI 与数字产品开发培训——四个月后大概又要再来一轮。
  • 对工程师角色的描述:编码助手让工程师把时间从「写代码」转向「指挥模型、验证输出、做更高阶决策」——结果是脑力负担更重,因为重复任务被剥离、留下批判性思考。
  • 「无悔之举」(no-regrets move)是为自己与组织投资学习与再学习——理解哪些工具对自己角色重要,如何指导落地,如何对输出做判断。

完整译文(中文)

MDDI 英文原文译文 · 翻译日期:2026-05-02

各位早安。

「AI」与「数字转型」我们听得太多了。我相信在座许多人已经领会到它们的重要性——所以我不会再花时间说服各位。

我想谈三个领导力挑战——这些挑战在我与「在 AI 时代领导组织」的同行交谈时反复出现。

把 AI 用得好,不是技术练习——这要求领导者付出实在的个人投入去推动组织变革,并以清晰、勇气和稳健的判断力去做这件事。

我们最常遇到的第一个挑战是——领导者要足够深入地理解自己的组织,才能知道 AI 在哪里能创造「真实价值」,而不只是「增量效率」。

今天许多领导者已经鼓励员工用 AI 来改进既有流程、或者提升生产力——比如更快整理报告、加速重复任务。这些努力确实提升了生产力——但 AI 的潜能远不止于此。

组织从 AI 或数字转型中提取不到更多价值的一个原因是——领导者也许不够了解组织里详细的工作流与流程。许多领导者花了多年精力在「升职阶梯、利益相关方互动、公关」上——现在我们必须回到组织里,理解这些流程与工作流多年来是怎么演化的。

我用我在卫生部的一段经历举个例子。2023 年走出疫情时,我们面临护士短缺,但搞不清原因。深入挖下去之后我们发现——我们的护士注册流程居然有 100 多个步骤!我们看着整个流程说:「这要做得更快」——而要点不是把 100 步「数字化」,而是先想清楚——「这 100 步是不是都需要?」我相信你们在各自的组织里也会面临类似的挑战。

起点是回到基本面,理解流程是怎样演化过来的——作为领导者,要有信念与勇气说一句「让我们从零开始」。

贯穿这门课程的过程中,是个好时机问自己:你是否真正理解组织里最迫切的问题——以及哪些是数字转型或 AI 能从根本上改变结果的领域?你准备好说「让我们从零开始」了吗?

领导者面临的第二个挑战是——一旦你搞清楚问题之后,要评估组织的当前能力,决定要「内建」(build)、「借用」(borrow)还是「购买」(buy)这些能力。

我自己也花了一些时间学 AI 的基础。我也很幸运,之前在公共部门工作时参与过几次 IT 系统升级。但这是枯燥而辛苦的工作。技术发展的节奏,又比我们学习的能力快得多。

没有一种方案对所有组织都适用。如果组织决定依赖外部供应商——你就要承担一些「员工去技能化」的风险,可能流失多年来熟悉这些流程的资深员工,丧失监督 AI 系统所需的领域专长。如果你尝试把所有能力都放在内部——转型节奏可能比你期望的更慢,变革的惯性会非常高。

这些都是领导者必须穿行其中的真实张力——决定变革的节奏、采用的深度,并与团队就「新的现实」进行坦诚的对话。

我们多数人最终走的是「混合方案」——引入一些外部能力,同时说服团队向更懂的人学习。我希望各位自己探索——你给自家组织的「混合配比」是什么?

但你必须支持自己的人走过这次转型。你必须看到——他们会焦虑——你必须为他们的能力投资,而不是奔向「空洞的效率」。

第三个挑战——也许是最重要的——是理解 AI 与数字转型对「人」的影响,并知道「即便能用,何时不该用 AI」。

并不是组织里的每一个决策、每一道流程或工作流,都该纯粹为了速度、规模或成本来优化。

作为领导者,你必须问自己:

这次 AI 的使用,会不会侵蚀「信任」?

它是否在共情与情感重要的场合,削弱了人的判断?

它是否让我们这些领导者远离了「问责」?

时不时问问自己——什么时候应该把 AI 挡在自家组织门外?因为人比效率更重要。

这就是「领导力」最关键的地方。AI 不自带道德罗盘,但你们都自带。我鼓励大家在这门课程中思考自己的罗盘——并让它在课程之后引导你的领导决策。

对领导者而言,一个明确的「无悔之举」是——为自己与组织投资学习与再学习。我们都需要学如何与 AI 共事——理解哪些工具对自身角色最关键、如何有效引导工具的落地、如何对其输出做合理判断。

我用软件工程举个例子——伴随编码助手的兴起,今天许多工程师(包括 GovTech 的工程师)花在「实际写代码」上的时间在减少,更多时间用于指挥 AI 模型、验证输出、做更高阶决策。

许多工程师告诉我——一开始他们曾担心工作会不会被取代、在校所学是否还相关。但现在他们的工作其实更费脑力——因为重心从「常规重复」转移到了「批判性思考」。

这种模式很可能会在许多职业与行业里复现。

因此——像我们这样的领导者必须保持「上手参与」、培育终身学习的心态、深入思考 AI 如何重新塑造我们的工作与责任。

新加坡的部长们去年也接受了 AI 与数字产品开发的培训——去年底我们做了两次。坦白说,再过四个月,我们大概又要再来一轮!

我很高兴出席新加坡国立大学(NUS)「AI 与数字转型高管硕士课程」的启动仪式。致首届班的所有同学——祝贺你们迈出这一步。在工作、生活与学业之间取得平衡,是对「在变化的世界中担当领导」的严肃承诺;这同时也是一种幸运——我们当中许多人都希望自己有同样的时间与资源,在如此结构化的环境中、与世界各行各业的同学一起学习。希望各位充分利用这次机会。

我也希望这门课程不仅装备你们以技术技能与知识,更装备你们以判断力、信心与伦理上的清晰——好把转型领导得漂亮。

感谢新加坡国立大学开启这一重要倡议——祝各位前路真正具有「转型意义」。

谢谢。

英文原文

MDDI 官网原始记录 · 抓取日期:2026-05-02

Good morning, everyone.

We hear the words “AI” and “digital transformation” all the time. I’m sure many of you already appreciate their importance, so I will not spend more time convincing you of that.

Instead, let me share three leadership challenges that repeatedly surface in my conversations with leaders who are leading their organisations through this AI age.

Using AI well is not a technical exercise. It requires significant personal commitment from leaders to drive organisational change. And to do so with clarity, courage, and sound judgment.

The first challenge we often encounter is for leaders to understand your organisations deeply enough to know where AI can create real value, not just incremental efficiency.

Today, many leaders already encourage their staff to use AI to improve existing processes or to become more productive, in organising reports or speeding up routine tasks. These efforts do improve productivity, but AI has the potential to do so much more.

One reason why organisations cannot seem to extract more value from AI or digital transformation is because leaders may not know well, the detailed workflows and processes being used within the whole organisation. Leaders may have spent years raising the ranks, stakeholder engagement, public relations etc, and we now need to go back into our organisations to understand how processes and workflows have evolved over time.

Let me give you a good example of the experience I had in the Ministry of Health. Coming out of COVID in 2023, we had a nursing shortage, but we couldn't figure out why. When we started digging into it, we found that our nursing registration process had over 100 steps in it. We looked at the whole process and said, we think we need to do this faster, and it was not about digitalising the 100 steps. It was to first think about if you need all of these steps. I think this is the kind of challenge you will face in your respective organizations.

It starts with going back to the basics, understanding how processes have evolved over time, and as a leader, having the conviction and the courage to say, let's start from zero.

As you go through this programme, it is a good time to ask yourselves: Do you really understand your organisation’s most pressing problems, and the areas where digital transformation or AI can fundamentally change outcomes? Are you ready to say, let's start from zero?

The second challenge leaders often face, once you figure out what the problems are, is assessing your organisation’s current capabilities to carry out the transformation, and then deciding whether you want to build in-house, borrow new capabilities or buy these capabilities.

It took me some time to learn basics of AI. I was also fortunate to have had to work through some IT system enhancements during my time in the public service. But this is dry and hard work. And the pace of technology development is much faster than our ability to learn.

There is no one size fits all solution for all organisations. If your organisation decides to rely on external vendors, you will risk some deskilling of your workforce, lose experienced staff who worked on the processes over the years, and lose domain expertise needed to supervise your AI systems. If you try to build all your capabilities in-house, the pace of transformation may be slower than you want, and the inertia to change will be very high.

These are real tensions that leaders must navigate. Deciding the pace of change, the depth of adoption, and having honest conversations with your teams about new realities.

Now, most of us end up on a hybrid – we bring in some external capabilities, and then we convince our teams to learn from those who know better. I would like you to explore on your own, what's the mix that you bring to your organisation?

But you will have to support your people through the transformation. You have to recognise that they will be anxious, and you have to invest in their capabilities, rather than hollow efficiency.

The third challenge, and perhaps the most important one, is understanding the impact of AI and digital transformation on people, and knowing when not to use AI, even when you can.

Not every decision, process or workflow in your organisation should be optimised purely for speed, scale, or cost.

As leaders, you will have to ask yourselves:

Does this use of AI erode trust?

Does it reduce human judgment where empathy and feelings matter?

Does it distance us as leaders from accountability?

Now and then, you ask yourself, when do you keep AI out of your organisation? Because the humans matter more.

This is where leadership matters most. AI does not come with a built-in moral compass, but all of you do. I encourage all of you to think about your own compass as you go through this course, and let it guide you through your leadership decisions after the course.

One clear “no-regrets” move for leaders is to invest in learning and relearning, for yourselves and for your organisations. We all need to learn how to work with AI: to understand which tools matter for our roles, how to guide the implementation of the tools effectively, and how to exercise sound judgment over their outputs.

Let me give you an example on software engineering. Nowadays, with the rise of coding assistants, many of our engineers, including those in GovTech, spend less time on actual coding tasks and more time directing AI models, validating outputs, and making higher-order decisions.

Many engineers tell me that at the start, they were wondering whether their job will be replaced and whether what they learned in school is no longer relevant. But now their current work is a lot more mentally demanding, because the focus has shifted to critical thinking, rather than routine and repetitive tasks.

This pattern will likely repeat across many professions and sectors.

That is why leaders like all of us must stay hands-on, cultivate a lifelong learning mindset, and think deeply about how AI reshapes our work and responsibility.

Even our ministers in Singapore have undergone training in AI and digital product development. We had our two training sessions towards the end of last year, and to be honest, I think in four months’ time, we probably need another round of these training sessions!

I am happy to be here with you today at the launch of NUS’s Executive Master in AI and Digital Transformation. To all of you in the inaugural cohort, congratulations on taking this step. Balancing your work, life, and studies is a serious commitment to leadership in a changing world. But it's also a blessing, as there are many of us here who wish that we had the same time and the resources to spend on learning in a very structured environment and with students from different industries all over the world. I hope that all of you make the most of it.

I also hope that this programme equips you not just with technical skills and knowledge, but with the judgment, confidence, and ethical clarity to lead transformation well.

I thank NUS for this important initiative, and I wish you all a truly transformative journey ahead.

Thank you.